HAL Id: hal-00259490https://hal.archives-ouvertes.fr/hal-00259490

Preprint submitted on 28 Feb 2008

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Origin of relative marking in HausaMahamane L. Abdoulaye

To cite this version:

Mahamane L. Abdoulaye. Origin of relative marking in Hausa. 2008. �hal-00259490�

Origin of relative marking in Hausa*

Mahamane L. AbdoulayeAbdou Moumouni University, Niamey, Niger

Abstract

Although the relative marking has been well described for a number of African languages, the issue of its possible sources has so far not been systematically addressed. This paper gives an account of the origin of the relative marking in Hausa and shows that for this purpose, it is necessary to leave aside the typical contexts with which it is usually associated, that is, relative clauses and out-of-focus clauses of constituent focus and wh-question constructions. Instead, this paper shows that - kèe and -kà, in their capacity as Relative Imperfective and Relative Perfective markers, first appeared in scene-setting clauses, i.e., adverbial clauses that carry presupposed information setting the event of the main clause. The paper proposes that the relative marking arose after periphrastic scene-setting clauses of the structure "as it is + clause" underwent clause merger to derive reduced scene-setting clauses with the structure "as + clause". Using grammaticalization theory, the paper, on the one hand, describes some systematic formal constraints on the merger process and, on the other hand, shows that four different types of scene-setting clauses have each a specialized function. These functions, which are the basic information function, expression of speaker-based inference, conveying speaker expressivity, and profiling (emphasis, contrast, and information request), can be cast in an overall grammaticalization scheme involving the scene-setting clauses and their derived constructions, i.e., relative and out-of-focus clauses.

1. IntroductionIn the literature on African languages, the relative marking is characterized as a set of special tense/aspect paradigms used in pragmatically marked constructions and, sometimes, in narrative storyline clauses (see among others Schachter 1973, Hyman and Watters 1984, Bearth 1993). In Hausa, the Relative Perfective and Relative Imperfective are thought to occur mainly in perfective and imperfective clauses when some constituent is relativized, focus-fronted or questioned, and also, for the Relative Perfective, in narrative contexts (cf. Caron 1991: 159f, 170f, Green and Reintges 2003, Jaggar 2001: 161ff, Newman 2000: 567ff, Wolff 1993: 423f, 427f, etc.; cf. note 5 for the complete paradigms). However,

1

as discussed in Abdoulaye (1992: 53f, 75n2, 1997, 2007a), Hausa also uses the relative marking in scene-setting clauses (SSCs), i.e., adverbial clauses that express background/presupposed information and which cannot be assimilated to relative clauses, out-of-focus clauses, or the narrative storyline. The SSCs are illustrated next:

(1) a. [Dà su-kèe tsòoro-n à kaamàa su], sun gudù.as 3p-RI fear-of imp.SBJ arrest 3p 3p.CPL flee'[As they are afraid of being arrested], they fled.'

b. [Dà su-kà gàji], sun koomàa inuwàa su-nàa huutàawaa.DA 3p-RP tire 3p.CPL return shade 3p-IPV resting'[As they are tired], they went under shade to rest.'

c. [Dà zaa à mayaɍ dà suu Tàawa], sun faarà sayaɍ dà kaaya-n-sù.as FUT imp move 3p Tawa 3p.PERF begin sell stuff-of-3p'As they will be reassigned in Tawa, they have started selling their things.'

(2) a. Su-nàa tsòoro-n à kaamàa su.3p-IPV fear-of imp.SBJ arrest 3p‘They are afraid of being arrested.’

b. Sun gàji.3p.CPL tire‘They are tired.’

c. Zaa à mayaɍ dà suu Tàawa.FUT imp move 3p Tawa‘they are going to be reassigned in Tawa.’

The sentences in (1) all contain an adverbial clause (in brackets) expressing a presupposed cause for the event in the main clause, as indicated in the translations. The adverbial clause is presupposed in the sense that its content is known or taken for granted by speech participants. In the adverbial clause in (1a), the highest verb appears in the Relative Imperfective su-kèe ‘3p-RI’, which contrasts with the general pragmatically neutral Imperfective su-nàa ‘3p-IPV’, as illustrated in (2a). In the adverbial clause in (1b), the verb bears the Relative Perfective su-kà ‘3p-RP’, which contrasts with the general pragmatically neutral Completive sun ‘3p.CPL’, as illustrated in (2b). Finally, the adverbial clause in (1c) contains the Hausa Future I (or zaa Future), which developed from the motion verb zâa ‘be going’ (cf. Abdoulaye 2001). It may be noted that the Future I has no relative form and

2

appears under the same form in pragmatically neutral contexts, as illustrated in (2c). The Relative Perfective, the Relative Imperfective, and the Future I are the only TAMs that can appear in the type of SSCs illustrated in (1). However, Hausa also has another type of SSCs, which are periphrastic near-equivalents to the SSCs in (1) and which do not use the relative marking. These SSCs are illustrated in the following (cf. also the examples given in Newman 2000: 505):

(3) a. [Dà ya-kè su-nàa tsòoro-n à kaamàa su], sun gudù.as it-be 3p-IPV fear-of imp.SBJ arrest 3p 3p.CPL flee'[As it is the case they are afraid of being arrested], they fled.'

b. [Dà ya-kè sun gàji], sun koomàa inuwàa su-nàa huutàawaa.as it-be 3p.CPL tire 3p.CPL return shade 3p-IPV resting'[As it is the case they are tired], they went under shade to rest.'

c. [Dà ya-kè zaa à mayaɍ dà suu Tàawa], sun faarà sayaɍ dàas it-be FUT imp move 3p Tawa 3p.CPL begin sellkaaya-n-sù.stuff-of-3p'As it is the case they will be reassigned in Tawa, they have started selling theirthings.'

The sentences in (3) have the same general meaning as those in (1). The SSCs in (3) differ from those in (1) in having an invariable impersonal copular predicate following the subordinator dà. The copula -kè ‘be’ takes as complement the adverbial clause proper, which, as indicated, takes the pragmatically neutral tense/aspect paradigms, i.e., the general Imperfective in (3a), the general Completive in (3b), and the Future I in (3c). Besides these three TAMs, the complement clause also allows other tense/aspect paradigms and nonverbal predicates, as will be seen in due course. This makes the periphrastic “copular” SSCs in (3) more general than the “reduced” SSCs given in (1). For this reason, Abdoulaye (1997, 2007a) suggests that reduced SSCs are derived from copular SSCs through clause merger, i.e., the merger between the copular ya-kè 'it-is' predication and the inner TAM of the adverbial clause. When the inner clause is in the Completive or Imperfective, the merger gives rise to the relative marking.

The aim of this paper is to present some evidence in support of the proposed clause merger analysis. To this effect, the paper uses grammaticalization theory. Normally, a grammaticalization process, in a specific context, turns a lexical or derivational item into a

3

grammatical marker, or a grammatical marker becomes more grammatical (cf. for example Hopper and Traugott 1994: 2). In this process, the original lexical item becomes progressively eroded, both at the phonological and semantic levels. However, there is another process that also falls under the domain of grammaticalization, where an entire construction undergoes expansion in new contexts and/or acquires new functions while becoming syntactically more integrated and less flexible. This type of grammaticalization has been discussed in, among others, Givón (1990: 651), Güldemann (2003: 183), Heine and Reh (1983: 34), Himmelmann (1997), and Hopper and Traugott (1994: 167ff). Indeed, this paper shows that changes in SSCs can be tracked on two levels, the functional and the formal levels. On the functional level, we will see that copular SSCs are primary and most naturally express the backgrounding of causal/reason clauses to assert the information in the main consequent clause. Reduced SSCs by contrast seem to add marked pragmatic and expressive features to their main clause. This is significant because the paper proposes that SSCs are built on temporal clauses introduced by the subordinator dà ‘when’. Indeed, a change from temporal markers to causal markers is well attested in world languages (cf. among others Hopper and Traugott 1994: 74f, Thompson and Longacre 1985: 181). On the formal level, copular SSCs are more complex and distributionally more general and have reduced counterparts only in certain contexts. It should be noted at the outset that the paper does not claim that the Relative Perfective and Relative Imperfective found in reduced SSCs are direct outputs of a morphosyntactic fusion between the copula kèe and the lower TAM. Rather, we will see that Hausa may have simply put into alternation the copula kèe with functionally equivalent markers (in particular, the Relative Perfective marker -kà) that existed independently in the history of Hausa.

The paper is organized as follows. Section 2 shows how basic copular SSCs are formed. Section 3 focuses on the formal aspect of SSCs and presents various conditions and constraints on the clause merger process. Section 4 shows that four types of SSCs have four different functions that can be better understood once set in an overall grammaticalization scheme implicating the SSCs and their derived constructions, i.e., relative and out-of-focus clauses.

2. Formation of copular SSCsThis section claims that copular SSCs are formed by the combination of an impersonal predicate ya-kè ‘it is’ and the temporal dà ‘when’ clause. The primary function of the new clause is to express a causal/reason event which serves as background to the main consequent event.

4

According to Newman (1976: 177), early descriptions of Hausa usually claim that the element kèe in central/eastern dialects appears in relative, focus and wh-question constructions, either as an auxiliarized Relative Imperfective marker -kèe or as a copula kèe/-kè in nonverbal predications (however, the copula, too, was most often wrongly referred to as "relative imperfective/ continuous"; cf. Schuh 2001 for a review). In both cases, kèe alternates with –nàa, the copula or auxiliary found in assertive contexts, as illustrated next using focalization in a verbal and a nonverbal (locative) predication: 1

(4) a. Yâara su-nàa wargii dà saabin kaayan wàasa-n-sù.children 3p-IPV play with new toys-of-3p'The children are playing with their new toys.'

b. Yâara (nèe) su-kèe wargii dà saabin kaayan wàasa-n-sù.children cop. 3p-RI play with new toys-of-3p'It is the children who are playing with their new toys.'

(5) a. Abdù ya-nàa nan.Abdu 3ms-be.at there'Abdu is around.'

b. Abdù (nee) kèe nan.Abdu cop. be.at there'It is Abdu who is around.'

In (4), both sentences are in the Imperfective, with –nàa marking the general (assertive) Imperfective in (4a) and –kèe marking the Relative Imperfective in the out-of-focus clause in (4b). The alternation in (4) is naturally of the same type as that illustrated in SSCs (cf. data (1a) and (2a) above). In (5), kèe and –nàa alternate as locative copulas in a nonverbal predication, again with kèe marking the non-assertive part of the predication. However, as clearly shown in Newman (1976: 181f), this account of kèe turned out to be too narrow since there are contexts where kèe does not alternate with –nàa in focus, relative, or wh-question constructions. These contexts are illustrated next:

(6) a. Àkuyà-ɍ (cee) kèe bàbba.goat-df cop. be big'It is the goat that is big.'

5

b. Àkuyà-ɍ bàbba cèe.goat-df big cop.'The goat is big.'

(7) a. Màalàmai ukkù (nee) kèe àkwai.teachers three cop. be.at there.is'It is three teachers that there is.'

b. Àkwai màalàmai ukkù.there.are teachers three'There are three teachers.'

(8) a. Ƙarfèe 2 nèe ya-kè sun zoo.o'clock two cop. it-be 3p.CPL comeIt is by 2 o'clock that they have arrived.'

b. Ƙarfèe 2 sun zoo.o'clock two 3p.CPL come'By 2 o'clock they have arrived.'

In data (6a), kèe marks the out-of-focus part of a focalized adjectival predication. The neutral predication uses the copula nee/cee, as seen in (6b), which establishes an alternation between kèe and nee/cee (note that the cee in (6a) is a focus marker). In (7a) kèe marks the out-of-focus part of an existential predication. In the neutral sentence (7b), one finds the existential predicate àkwai ‘there.is/are’ alone. Sentence (8a) illustrates focalization in an anterior construction. In this case, the normal Relative Perfective cannot be used since, as a true perfective, it would eliminate the anterior meaning (cf. ƙarfèe biyu (nèe) sukà zoo 'it is at 2 o'clock that they arrived). Instead, there seems to be the insertion of an impersonal predication ya-kè ‘it-be’, which takes the rest of the clause as complement in an equational construction, and so marks it as the non-asserted part of the sentence (compare with the neutral (8b)). In (7-8) then, kèe, strictly speaking, alternates with nothing. Nonetheless, the illustrations so far given do not exhaust the distribution of kèe and indeed some other uses are reported in the literature but are either quickly discarded as improductive remnants (cf. Newman 1976: 178n3) or they receive an inadequate characterization (cf. Jaggar 2001: 177f, 463f; cf. also note 2). The main feature of these uses is that kèe cannot be associated with the typical Hausa relative, focus, or wh-question construction. One of these uses is illustrated next (cf. also Abdoulaye 2006: 1141f):

6

(9) a. Bàlki ta-nàa dà wàayoo.Balki 3fs-have cleverness'Balki is clever.'

b. Bàlki (cèe) ta-kèe dà wàayoo./ Wàayoo (nèe) Bàlki ta-kèe dà shii.Balki cop. 3fs-have cleverness/ cleverness cop. Balki 3fs-have 3ms'It is Balki who is clever./ 'It is clever that Balki is.'

c. Bàlki dà wàayoo ta-kè.Balki wih cleverness 3fs-be'Balki is really clever.'

Example (9a) presents a pragmatically neutral possessive construction (with a characterizing function). The two arguments in the predication can each be focalized in the typical way, with complete fronting and the focus marker nee/cee, as indicated in (9b). However, as seen in (9c), there exists another pragmatically marked sentence where the second argument (in the neutral predication) obligatorily appears just before the copula kèe, and the construction has an emphatic meaning, as indicated. In (9c), none of the two NPs is focus-fronted (and the focus marker nee/cee cannot be used) so that kèe here is not associated with a relative, focus, or wh-question structure. In fact, there is even a clearer case showing the independence of kèe vis-à-vis focus and assimilated constructions, as seen next (cf. also Abdoulaye 2007a): 2

(10) a. (Wannàn) mài gaadìi kèe nan.this watchman be.at there‘This is the watchman (whom I just talked to you about).’'This is the watchman (you expect there is one for this place).'

b. (Wannàn) mài gaadìi nee.this watchman cop‘This is the/a watchman.'

In (10a), kèe is combined with a bleached locative demonstrative nan ‘there’ and the whole complex has a deictic identification function, as indicated. Although (10a) has no focus structure or meaning, it does have a certain pragmatic function. Indeed, in (10a) kèe nan is a presuppositional identification copula that typically indicates that the hearer already heard or has information about the identified referent (first interpretation), or that he/she expects given the context (second interpretation). By contrast, sentence (10b), with the general

7

identificational copula nee/cee, identifies a new referent to the hearer in a pragmatically neutral context. Sentence (10b) is typically followed by further information about the referent.

Data (9-10) clearly establish kèe (or its variant –kè with a prefixed pronoun) as a presupposition marker, whether or not it is associated with a grammatically marked focus, relative, or wh-question construction. This characterization of kèe is well supported by its functions in the two main Hausa dialect clusters, the west and the central/eastern dialects, as represented in Table 1.

Table 1: Some functions of kèe in Hausa dialects

West dialects Central/east dialectsPresuppositional identification copula kèe nan kèe nanPresupposition marker in emphatic possession -kè -kèPresupposition marker in anterior focus -kè -kèPresupposition marker in focus of nonverbal complement/ locative NPs

kèe/-kè(e) kèe/-kè(e)

Presupposition marker in focus of nonverbal subjects

aɍ kèe

Presupposition marker in constituent focus in Imperfective clauses

-kà -kèe

In Table 1, one notes that only certain functions of kèe are pandialectal, i.e., the presuppositional identification function and presupposition marking in marginal (and probably archaic) focus/emphatic constructions. In typical constituent focus constructions in Imperfective clauses, the western dialects do not use kèe, but a marker -kà (cf. Caron 1991: 23, 126f, 147; cf. note 5 on the history of this marker). This shows that the use in focus, relative, and wh-question contexts maybe is secondary and that kèe, after it was driven out of assertive contexts by newer elements (i.e., -nàa, nee/cee, existential dà and àkwai, etc.), was first of all relegated to the function of a presupposition marker.

It is no surprise then when kèe (or, more precisely, the impersonal locution ya-kè ‘it-be’) is combined with the temporal marker dà ‘when’ to form presupposed causal/reason adverbial clauses meaning ‘when [as] it is the case...’, i.e., the copular SSCs, which are not relative or out-of-focus clauses. This is illustrated in the following: 3

8

(11) a. [Dà bàaƙii sun fìta], kà rufè ƙoofàa.when guests 3p.CPL exit 2ms-SBJ close door'As soon as the guests go out, you may close the door.'

b. [Dà ya-kè bàaƙii sun fìta], sai kà rufè ƙoofàa.as it-be guests 3p.CPL exit then 2ms-SBJ close door'As it is the case the guests went out, you may now close the door.'

(12) a. [Dà su-nàa aikì-n yinìi], a-nàa kaawoo ma-sù àbinci.when 3p-IPV work-of day imp-IPV bring to-3p food'When they were working all day, they were supplied with food.'

b. [Dà ya-kè su-nàa aikì-n yinìi], a-nàa kaawoo ma-sù àbinci.as it-be 3p-IPV work-of day imp-IPV bring to-3p food'As it is the case they are working all day, they are supplied with food.'

(13) a. [Dà zaa à mayaɍ dà ita Tàawa], an ƙaaràa ma-tà àlbâashii.when FUT imp move 3fs Tawa imp.CPL raise to-3fs salary'When she was about to be reassigned in Tawa, her salary was raised.'

b. [Dà ya-kè zaa à mayaɍ dà ita Tàawa], an ƙaaràa ma-tà àlbâashii.as it-be FUT imp move 3fs Tawa imp.CPL raise to-3fs salary'As it is the case she is going to be reassigned in Tawa, her salary was raised.'

Sentences (11a), (12a), and (13a) contain simple temporal adverbial clauses introduced by dà ‘when’. As suggested in Abdoulaye (2007b), this temporal subordinator may have derived from comitative and instrumental marker dà ‘with, and’. Sentences (11b), (12b), and (13b) on the other hand contain causal/reason adverbial clauses, as indicated in the translations. Another difference between the two sets of sentences is that the temporal clauses are in the assertion domain of their sentence, while the causal/reason clauses, because of the presence of ya-kè ‘it be’, carry presupposed information, which is known or taken for granted by the speech participants. In this paper, I will assume that the causal function of the adverbial clause is due to the subordinator dà, which can change from a temporal to a causal marker, particularly when it introduces an equational clause built with the copula ya-kè. As already suggested in Section 1, the subject in the causal clause is the impersonal ya- ‘it' (<3masc. sing.) and the complement the inner tensed clause, in a "as it is + clause" structure (for other functions of yakè in relative clauses, see Abdoulaye 1992: 75n2, 2007a, 2007b, McConvell 1977: 19ff, and Newman 2000: 540f). 4

9

To summarize, Hausa has a copula kèe whose basic function is to mark presupposed information. This copula, when it heads impersonal equational predications, can combine with the subordinator dà to form copular SSCs. The next section describes the clause merger process that derives reduced SSCs from copular SSCs.

3. Clause merger in Hausa SSCs

The central tenet of this paper is that the relative marking, as Perfective and Imperfective tense/aspect paradigms, arose first in reduced SSCs, following a clause merger in the source copular SSCs. This section presents four cases of such merger, including the two that gave rise to the relative marking. The section also reviews various kinds of formal constraints that can block the merger.

3.1 Four cases of clause reduction in SSCsThe first case of merger in copular SSCs involves nonverbal complement clause, i.e., equational, nominal, or adjectival predications, and locative or possessive predications. The nominal and possessive predications are illustrated next:

(14) a. [Dà ya-kè Bàlki ƴaɍ dàariyaa cèe], hiiɍâ-ɍ taa yi daaɗii.As it-be Balki laugh.person cop. chat-df 3fs.CPL do pleasure‘As it is the case Balki is an entertaining person, the chat was pleasant.’

b. [Dà Bàlki ta-kè ƴaɍ dàariyaa], (ai) hiiɍâ-ɍ taa yi daaɗii.As Balki 3fs-be laugh.person well chat-df 3fs.CPL do pleasure‘As Balki is an entertaining person, the chat was, naturally, pleasant.’

(15) a. [Dà ya-kè ta-nàa dà haalii], taa sàyi gidâ-n mài tsàadaa.as it-be 3fs-have wealth 3fs.CPL buy house-df expensive‘As it is the case she is rich, she bought the expensive house.’

b. [Dà ta-kèe dà haalii], (ai) taa sàyi gidâ-n mài tsàadaa.as 3fs-have wealth well 3fs.CPL buy house-df expensive‘As she is rich, she, naturally, bought the expensive house.’

In the adverbial clause of (14a), the predicate ya-kè ‘it-is’ takes as complement an equational clause containing the regular equational copula nee/cee. In (14b), the adverbial clause is simplified in a subject-to-subject raising-like process. As a result, in (14b) the subordinator dà introduces a simple equational predication, with Bàlki as subject of the copula -kè. Data (15) shows the same process involving a possessive ‘have’ clause as

10

complement of ya-kè ‘it-is’. As stated in the introductory section, this paper does not assume that (14b) and (15b) are outputs of a rule of syntactic fusion in the copular SSC (for example, for (14b), application of a subject-to-subject raising and deletion of the neutral copula nee/cee, etc.). Rather, as seen in the previous section, the copula kèe exists in a variety of contexts with no relation to the SSCs.

The second case of clause merger in copular SSCs happens in imperfective context, as illustrated next:

(16) a. [Dà ya-kè Aishà ta-nàa sôn shubkà gyàɗaa], ta-nàa saaran daajìi.as it-be Aishà 3fs-IPV want plant peanuts 3fs-IPV cut bush'As it is the case Aisha wants to plant peanuts, she clears the bush.'

b. [Dà Aishà ta-kèe sôn shubkà gyàɗaa], (ai) ta-nàa saaran daajìi.as Aisha 3fs-RI want plant peanuts well 3fs-IPV cut bush'As Aisha wants to plant peanuts, she naturally clears the bush.'

In the adverbial clause of (16a), the predicate ya-kè ‘it-is’ takes a verbal clause as complement, a clause that carries the general Imperfective auxiliary –nàa. In the reduced SSC in (16b), the ya-kè predicate seems to have been dropped, but the remaining clause now has the Relative Imperfective marked by –kèe, instead of -nàa. This paper will assume that the Relative Imperfective auxiliary –kèe in (16b) developed from the locative copula kèe just as auxiliary –nàa developed from locative copula –nàa. One, therefore, need not assume that the copula in the ya-kè complex mechanically shifted into the complement clause. Rather, the sequence "ya-kè + regular Imperfective complement clause" can functionally be replaced (or can alternate) with an independently existing Imperfective clause built with the auxiliary –kèe in the context of causal/reason dà.

The third case of clause merger in copular SSCs happens when the complement clause of ya-kè ‘it-is’ has the Completive TAM. To this configuration corresponds in the reduced SSC an element -kà, as illustrated next:

(17) a. Dà ya-kè sun cêe su-nàa zuwàa, mun yi jirà-n-sù.as it-be 3p.CPL say 3p-IPV come 1p.CPL do wait-of-3p'As it is the case they advised us they are coming, we waited for them.'

b. Dà su-kà cêe su-nàa zuwàa, (ai) mun yi jirà-n-sù.as 3p.RP say 3p-IPV come well 1p.CPL do wait-of-3p'As they advised us they are coming, we, of course, waited for them.'

11

In (17a) the copula -kè takes as complement a clause whose highest verb cêe ‘say’ carries the Completive TAM sun ‘3p.CPL’, which has a perfect meaning (cf. Newman 2000: 569ff, Caron 1991: 164ff, and Schubert 1971/72: 220f). The reduced SSC in (17b) is simplified and the subordinator dà is directly followed by the verb cêe ‘say’. The clause carries the Relative Perfective su-kà ‘3p-RP’, which has a true perfective meaning (cf. Abdoulaye 2007c). As in the previous cases, the marker -kà may not be a direct morphological output of a merger process. Rather, it is more likely that -kà, before being recruited as the Relative Perfective marker (in central/eastern dialects), existed in the language, probably with a non-aspectual function. 5

The last case of clause merger in copular SSCs is observed when the clause complement of ya-kè carries the Hausa Future I or zaa-Future, as illustrated next:

(18) a. [Dà ya-kè zaa à mayaɍ dà suu Tàawa], sun faarà sayaɍ dàas it-be FUT imp move 3p Tawa 3p.CPL begin sellkaaya-n-sù.stuff-of-3p'As it is the case they will be reassigned in Tawa, they have started selling theirthings.'

b. [Dà zaa à mayaɍ dà suu Tàawa], sun faarà sayaɍ dà kaaya-n-sù.as FUT imp move 3p Tawa 3p.PERF begin sell stuff-of-3p'As they will be reassigned to Tawa, they have started selling their things.'

In (18a), the copula -kè takes as complement a clause containing the Future I. In the reduced version in (18b), only the ya-kè complex seems to have been left out, since the inner TAM does not change its shape, i.e., to the Future I corresponds no special relative form. It should be noted that negative nonverbal predications and negative Completive, Imperfective and Future I clauses all allow the merger but use no special relative form (cf. Abdoulaye 1997: 316n3).

To summarize, one observes in Hausa a correspondence between semantically similar copular SSCs and reduced SSCs in at least four distinct cases. Two of these cases give rise to the relative marking, i.e., special Perfective and Imperfective markers that alternate respectively with the general Completive and the general Imperfective. We have concluded that in the four cases, there is reduction or clause merger from copular SSCs to reduced SSCs, even though the derived SSCs in fact make use of independently available markers, in

12

particular the relative marker -kà. The rest of this section provides further arguments for the clause merger analysis.

3.2 Tense/aspect conditions on reduction in SSCsIn most languages with relative marking, the alternation concerns only a few TAMs. We have seen in this section that in SSCs, Hausa allows reduction in only four cases. Indeed, TAMs other than the Completive, the general Imperfective and the Future I (and their negative forms) either do not allow the reduction or cannot even appear in the copular SSCs. Paradigms appearing in copular SSCs but not allowing reduction are the Future II and the Habitual. The Subjunctive, the Eventual, and the Imperative do not at all appear in SSCs. This is illustrated next (cf. Abdoulaye 1997: 318f):

(19) a. Dà ya-kèÀali yâa jee kàasuwaa, Bàlki taa baa shì sabtù.as it-be Ali 3ms.FUT go market Balki 3fs.CPL give 3ms shop'As it is the case Ali will go to the market, Balki asked him to shop for her.'

b. *Dà Àali yâa jee kàasuwaa, Bàlki taa baa shì sabtù.as Ali 3ms.FUT go market Balki 3fs.CPL give 3ms shop'As Ali will go to the market, Balki asked him to shop for her.'

(20) a. Dà ya-kè Àali ya-kàn jee kàasuwaa, ka-nàa iyà baa shì sabtù.as it-be Ali 3ms-HAB go market 2ms-IPV can give 3ms shop'As it is the case Ali usually goes to the market, you can ask him to shop for you.'

b. *Dà Àali ya-kàn jee kàasuwaa, ka-nàa iyà baa shì sabtù.as Ali 3ms-HAB go market 2ms-IPV can give 3ms shop'As Ali usually goes to the market, you can ask him to shop for you.'

(21) a. ??Dà ya-kè Àali ya-kàa yi wannàn aikìi, an tàmbàyee shì.as it-be Ali 3ms-EVE do this deed imp-CPL ask 3ms'As it is the case Ali may have done this deed, he was questioned.'

b. *Dà Àali ya-kàa yi wannàn aikìi, an tàmbàyee shì.as Ali 3ms-EVE do this deed imp-CPL ask 3ms'As Ali may have done this deed, he was questioned.'

In (19a) and (20a), the clause complement of ya-kè carries, respectively, the Future II and the Habitual in the copular SSC. As shown in (19b) and (20b), these SSCs cannot be reduced, i.e., the Future II and the Habitual cannot appear in reduced SSCs. Sentences (21)

13

show that the Eventual is unfelicitous in both types of SSCs. The pattern of interaction between the TAMs and the SSCs can be summarized as in Table 2.

Table 2: Interaction between TAMs and SSCs (Central/east dialects)

Copular SSC Reduced SSCNonverbal clauses Yes Yes (kèe)Imperfective Yes Yes (Rel. Imperf.)Completive Yes Yes (Rel. Perf.)Future I Yes YesFuture II Yes NoHabitual Yes NoEventual No NoSubjunctive No NoImperative No No

In Abdoulaye (1997), it is proposed that reduced SSCs, with the relative marking, describe presupposed and realis events. The Future I is able to appear in the realis environment of reduced SSCs possibly because it grammaticalized from the verb zâa 'start to go, be going', which has an inchoative or action in progress meaning (cf. Abdoulaye 2001). Copulas such as nee/cee in equational, nominal, and adjectival predications and -nàa in locative and possessive 'have' predications can be considered to describe realis state of affairs and hence satisfy the requirement for the reduced SSCs. It may be assumed that the ya-kè predicate, too, marks both presupposition and realis status, whether or not the complement clause describes a realis event (for example, in (19a), ya-kè predicates as true a situation where Ali is projected to go to the market next day). When the situation in the complement clause is itself realis, as in (14-18) above, reduction is possible since there is feature compatibility between predicate ya-kè and the inner TAM. In other words, in the context of adverbial modification, the TAM in the reduced SSC, whether or not it assumes a special form, expresses both presupposition and realis status. However, when the situation in the complement clause is not realis, as in (19a), then the merger is not possible, since there is feature incompatibility and the two clauses must be kept distinct. Furthermore, ya-kè apparently cannot predicate a situation that is neither realized nor firmly projected to be realized in the future, hence the inability of the Subjunctive and other pure irrealis/hypothetical TAMs to appear in copular SSCs. 6

14

To summarize, although Relative Imperfective marker -kèe and Relative Perfective marker -kà cannot be taken as direct outputs of syntactic or morphological fusion rules, in SSCs they participate in a pattern of interaction between predicate ya-kè 'it-is' and the TAMs of the complement clause and can hence be considered, in their new capacity as exponents of the relative marking, to stem out of a clause merger process.

3.3 Formal constraints on reduction in SSCsAnother indication supporting the clause merger analysis is the fact that certain operators and special syntactic configurations do not appear in reduced SSCs, even if the complement clause contains realis TAMs in the corresponding copular SSC. For example, if in a copular SSC the complement clause is part of a conditional construction and is preceded by in/ìdan 'if', then reduction is not possible, as illustrated next:

(22) a. Dà ya-kè ìdan sun tsayàa âa kaamàa su, sai su-kà wucèe.as it-be if 3p.CPL stop imp.FUT arrest 3p then 3p-RP pass'As it is the case if they stay they would be arrested, they continued.'

b. *Dà ìdan su-kà tsayàa âa kaamàa su, sai su-kà wucèe.as if 3p-RP stop one.FUT arrest 3p then 3p-RP pass'As if they stay they woul be arrested, they continued.'

In the copular SSC in (22a), the copula -kè takes as complement a conditional construction marked with in/ìdan 'if' and whose first TAM is the Completive sun '3p.CPL'. As indicated in (22b), there is no corresponding reduced SSC. This failure to reduce may be due to the fact that the conditional operator in/ìdan must itself be under the scope of the ya-kè 'it-is' predicate and in case of a reduction, the operator will necessarily precede the relative marking, i.e., it will be out of scope for the presupposition and realis features marked by the relative marking. The same pattern of interaction between in/ìdan and SSC reduction is also observed in nonverbal clauses (cf. dà yakè ìdan likitàɍ màce cèe... 'as it is the case if the doctor is a woman...', with no reduced form) or in Imperfective or Future I complement clauses. Besides in/ìdan 'if', other operators that block reduction in copular SSCs are koo 'even', haɍ 'till, even, etc.', sai 'till, unless, must, etc.', duk 'all', etc.

Similarly, if the complement clause of a copular SSC has a focus or a topicalized structure, then reduction is not possible, as seen next:

15

(23) a. Dà ya-kè ɗàrii nèe ta-kèe sôo, sai kuɗî-n bà sù isa ba.as it-be hundred cop. 3fs-RI want then money-df NEG.CPL 3p suffice NEG'As it is the case she wanted a hundred [something], the money wasn't enough.'

b. *Dà ɗàrii nèe ta-kèe sôo, sai kuɗî-n bà sù isa ba.as hundred cop.3fs-RI want then money-df NEG.CPL 3p suffice NEG'As she wanted a hundred [something], the money wasn't enough.'

(24) a. Dà ya-kè shii Abdù ya-nàa dà mootàa, sai ya kai sù can.as it-be 3ms Abdu 3ms-have car then 3ms.RP take 3p there'As it is the case, Abdu [as for Abdu] has a car, he took them there.'

b. *Dà shii Abdù ya-kèe dà mootàa, sai ya kai sù can.as 3ms Abdu 3ms-have car then 3ms.RP take 3p there'As Abdu [as for Abdu] has a car, he took them there.'

In (23a) and (24a), the complement clause in the copular SSC starts with a focalized or topicalized constituent, respectively. As shown in (23b) and (24b), reduction is blocked in both cases. Again, the blocking may be explained by assuming that in (23a) and (24a), every constituent in the complement clause is under the scope of the higher predicate ya-kè. In this perspective, focus or topicalized constituents will be out of the scope of the relative marking in a reduced SSC (as seen in (14b) and (16b), normal nominal subjects must be assumed to remain under the scope of the relative marking, maybe through the subject pronoun/agreement).

In Section 3.1, we saw that equational, nominal, and adjectival predications as well as locative and possessive 'have' predications allow the merger when they appear as complement in copular SSCs. It seems however that all other nonverbal clauses appearing as complement of ya-kè 'it-is' do not allow clause merger. Similarly, clauses containing the defective motion verb zâa 'start to go, be going' also seem to disallow the merger. The next data illustrates one-term identification, existential, and zâa 'go' predications:

(25) a. Dà ya-kè Bàlki cèe, sun yàɍda.as it-be Balki cop. 3p.CPL agree'As it is the case it is Balki, they agreed.'

b. *Dà Bàlki cèe/ ta-kè, sun yàɍda.as Balki cop./ 3fs-be 3p.CPL agree'As it is Balki, they agreed.'

16

(26) a. Dà ya-kè àkwai mutàanee à fiilî-n, an baɍ wutaa kùnne.as it-be there.are people in place-df imp.CPL leave light on'As it is the case there are people in the place, the light was left on.'

b. *Dà àkwai mutàanee à fiilî-n, an baɍ wutaa kùnne.as there.are people in place-df imp.CPL leave light on'As there are people in the place, the light was left on.'

(27) a. Dà ya-kè zâa ta Saafòo gòobe, an jinkìrtà aikì-n.as it-be go 3fs Safo tomorrow imp.CPL delay work-df'As it is the case she goes to Safo tomorrow, the work is delayed.'

b. *Dà zâa ta Saafòo gòobe, an jinkìrtà aikì-n.as go 3fs Safo tomorrow imp.CPL delay work-df'As she goes to Safo tomorrow, the work is delayed.'

In (25a), the copular SSC has a complement clause that consists in a one-term identificational predication (which is typically used for direct, visual identification; cf. Abdoulaye 2007a). Data (25b) shows that this predication does not allow clause merger., whether one uses the neutral copula nee/cee or the presuppositional -kè copula. Data (26) shows that existential predications built with àkwai 'there is/are' do not allow the clause merger (the same observation applies to the alternate existential predicate dà 'there is/are', but apparently not to the negative existential bâa/baabù, as seen in: dà yakè bâa koowaa... = dà bâa koowaa 'as it is the case there is nobody...'). In data (27), the verb zâa 'start to go, be going' is defective and in fact has no overt identifiable tense/aspect paradigm (cf. Abdoulaye 2001: 5). As in seen in (27b), it cannot appear in reduced SSCs. Overall, the difference between the nonverbal predications illustrated in (25-27) and those in (14-15) is that the former have only one argument. The exact reason why one-argument nonverbal predications would not allow reduction in SSCs will be left to future research.

Another formal constraint in the reduction of SSCs relates to the complexity of the inner complement clause. Indeed, it seems that for reduction to be possible, a temporal adverbial clause should not intervene between the ya-kè 'it-is' predicate and the highest verb of the complement clause. This is illustrated next:

(28) a. [Dà ya-kè [lookàcin dà ya-nàa aikìi] yaa ji zuwà-n-sù]...as it-be time that 3ms-IPV work 3ms.CPL hear coming-of-3p'[As it is the case that [when he was working] he heard them coming]...'

17

b. *[Dà [lookàcin dà ya-nàa aikìi] ya ji zuwà-n-sù]...'[As [when he was working] he heard them coming]...'

c. [Dà ya ji zuwà-n-sù [lookàcin dà ya-nàa aikìi]]...'[As he heard them coming [when he was working]]...'

Example (28a) presents a copular SSC (main clause left out) where the complement clause of ya-kè 'it-is' has a preposed temporal adverbial clause (cf. the inner brackets). In this configuration, reduction is not possible, as seen in (28b). When the temporal clause is shifted to the right in the complement clause, then reduction becomes possible, as seen in (28c). In fact, there may be a more general constraint to the effect that any TAM intervening between the predicate ya-kè 'it-is' and the highest TAM of the complement clause can block the merger. This is illustrated next, with inner subject arguments containing a clause:

(29) a. [Dà ya-kè [cêewaa su-nàa ganin sarkii] ya-nàa ban màamaakìi]...as it-be given 3p-IPV see emir 3ms-IPV give surprise'As it is the case that the fact that they see the emir is surprising...'

b. *[Dà [cêewaa su-nàa ganin sarkii] ya-kèe ban màamaakìi]...as given 3p-IPV see emir 3ms-RI give surprise'As the fact that they see the emir is surprising...'

(30) a. [Dà ya-kè [bàaƙi-n dà su-kà zoo] su-nàa sôn fìtaa]...as it-be guests-df that 3p-RP come 3p-IPV want go.out'As it is the case that the guests who came want to go out...'

b. ??[Dà [bàaƙi-n dà su-kà zoo] su-kèe sôn fìtaa]...as guests-df that 3p-RP come 3p-IPV want go.out'As the guests who came want to go out...'

(31) a. [Dà ya-kè [indà su-kà tàfi] ya-nàa dà niisaa]...as it-be where 3p-RP go 3ms-have distance'As it is the case [the place] where they went is far...'

b. *[Dà [indà su-kà tàfi] ya-kèe dà niisaa]...as where 3p-RP go 3ms-have distance'As [the place] where they went is far...'

Examples (29a) and (30a), respectively, portray a copular SSC (main clause left out) where the subject of the complement clause is a that-clause or contains a relative clause. As shown

18

in (29b) and (30b), in both cases the clause merger is not possible, although the grammaticality is not always as bad as in the case of an intervening adverbial clause, as illustrated in (28). The same blocking effect is observed in (31a) with a locative adverbial relative clause as inner subject. This also applies to copular SSCs containing a manner adverbial clause as inner subject of the complement clause (cf. dà yakè yaddà sukèe yîi yanàa dà ban wàhalàa... 'as it is the case that how [the way] they do it is tiresome...’, with no possible reduction).

Finally, it may be noted that in a complex complement clause, only the highest (and also the first) TAM of the clause is relevant for the merger and eventually reflects the relative tense/aspect alternation. This is illustrated next:

(32) a. [Dà ya-kè mun ji Abdù yaa cêe ta-nàa aikì-n-sù kàasuwaa]...as it-be 1p.CPL hear Abdu 3ms.CPL say 3fs-IPV send-of-3p market'As it is the case we heard Abdu say that she sends them to the market...'

b. [Dà mu-kà ji Abdù yaa cêe ta-nàa aikì-n-sù kàasuwaa]...as 1p-RP hear Abdu 3ms.CPL say 3fs-IPV send-of-3p market'As we heard Abdu say that she sends them to the market...'

In (32a), the portrayed copular SSC (main clause left out) has a complement clause that itself contains three clauses. As seen in (32b), only the highest TAM, mun '1p.CPL', is concerned by the merger and is replaced with the Relative Perfective mu-kà '1p-RP'. Replacing the other two TAMs with the relative marking would indeed lead to ungrammaticality.

To summarize, in the TAMs and nonverbal predications that allow the clause merger, the process is still subject to a number of constraints. We have seen that certain operators and syntactic configurations such as focus and topicalization prevent the merger. Similarly, nonverbal predications with one argument tend not to allow the merger. Finally, in case the complement clause of ya-kè is complex, then reduction is possible only with regard to the highest TAM and when no other TAM precedes it. These various constraints and the interactions with the TAMs show that there is a connection between copular SSCs and reduced SSCs and that the latter are indeed derived from the former through clause merger. The next section shows that functional considerations, too, point to a connection between copular and reduced SSCs.

19

4. Functional aspects of the grammaticalization of SSCsThe aim of this section is to show that the claim that reduced SSCs are derived from copular SSCs is also supported by the functional patterns of different types of SSCs. The section essentially shows that copular SSCs are associated with a function closer to the source temporal/causal clause, while reduced SSCs have drifted away from this source. We will also see that in a few contexts, independently of the formal constraints seen in preceding section, only copular SSCs are available.

So far in this paper, the SSCs illustrated are causal/reason adverbial clauses that provide a background to the event in the main consequent clause. In fact however, it is possible for both copular SSCs and reduced SSCs to background a causal/reason clause or a consequent clause and each of the four logical possibilities has a distinct function. Indeed, we will see that copular SSCs fulfill basic information function when they background the causal/reason clause but convey speaker-based inference when they background the consequent clause. The reduced SSCs on the other hand modulate speaker expressivity when they background causal/reason clauses and mark profiling when they background consequent clauses. The contrast causal vs. consequent clause backgrounding in copular SSCs is illustrated next (cf. also Abdoulaye 1997: 312):

(33) a. [Dà ya-kè Aishà ta-nàa sôn shubkà gyàɗaa], ta-nàa saaran daajìi.as it-be Aishà 3fs-IPV want plant peanuts 3fs-IPV cut bush'As it is the case Aisha wants to plant peanuts, she clears the bush.'

b. Dà ya-kèAishà ta-nàa saaran daajìi, ta-nàa sôn shubkàgyàɗaa (kèe nan).as it-be Aishà 3fs-IPV cut bush 3fs-IPV want plant peanuts this.means'As it is the case Aisha is clearing the bush, (this for me means) she wantsto plant peanuts.'

In (33a), where the causal/reason clause is backgrounded in the copular SSC, the speaker evokes two connected states of affairs, one of which (the causal event) is presupposed or taken for granted. By contrast, the consequent event in the main clause is asserted, whether as new information or as something that is being explained to the hearer. However, when the consequent clause is backgrounded in the copular SSC, as in (33b), then the speaker is using some evidence or a reported fact (the presupposed consequent situation) to infer its cause or reason. This is shown by the fact that (33b) is in fact more natural if it is completed with the inference marker kèe nan 'this for me means' (cf. note 2). It should be noted that the inference is speaker-based only and has no implication whether or not the hearer already

20

came to the same conclusion by himself/herself. In fact, (33b) can be used to submit the inference to the appreciation of a more knowledgeable hearer.

On the ground of the contrast "hearer information" vs. "speaker-based inference" alone maybe one can conclude that causal clause backgrounding is the primary function of the copular SSCs. However, there is another clearer indication for the basicness of the cause/reason backgrounding function. Indeed, it is not always possible, from a consequent situation, to infer a cause and, in this case, consequent clause backgrounding is not possible, as seen next:

(34) a. Dà ya-kèbandìɍ ɗàrii ta-kèe sôo, sai kuɗî-n bà sù ìsa ba.as it-be rolls hundred 3fs-IPV want then money-df NEG CPL suffice NEG'As it is the case she wanted hundred rolls, the money was not enough.'

b. ??Dà ya-kè kuɗî-n bà sù ìsa ba, bandìɍ ɗàrii ta-kèeas it-be money-df NEG CPL suffice NEG rolls hundred 3fs-IPVsôo (kèe nan).want this.means'As it is the case the money was not enough, (this for me means) she wanted hundred rolls.'

(35) a. Dà ya-kè yâu Lahàdi nèe, ma'àikàtaa su-nàa huutuu.as it-be today Sunday cop. workers 3p-IPV rest'As it is the case today is Sunday, workers are resting.'

b. ??Dà ya-kè ma'àikàtaa su-nàa huutuu, yâu Lahàdi nèe (kèe nan).as it-be workers 3p-IPV rest todaySunday cop. this.means'As it is the case workers are resting, (this for me means) today is Sunday.'

In (34a), the causal situation is the desire to acquire 100 rolls and the consequent situation that the money was not enough (i.e., the amount of money failed to buy the desired quantity of rolls). In sentence (34b), the consequent situation cannot be backgrounded since, normally, one cannot tell the precise quantity of rolls desired just from the fact that a certain amount of shopping money was insufficient. Similarly, in (35b) the consequent clause cannot be backgrounded since one cannot infer it is Sunday just from the observation that people are not working.

In Section 3, we have presented many couples of copular SSCs and reduced SSCs encoding a causal/reason clause and have given the two constructions as equivalent. In fact,

21

though copular and reduced SSCs have the same general meaning, they are not pragmatically equivalent. Reduced SSCs encoding a causal/reason clause indeed seem to convey speaker conviction, expressivity, attitudes, or emotions about the proposition in the main clause. In data (14-18) above, the translation of sentences with reduced SSCs contains the words “naturally” or “of course” to show that the main purpose of the speaker is to express a conviction, more than factual information. Reduced SSCs are also frequently associated with rhetorical style, irony/sarcasm, pride/honor, derision, and scolding. Irony or sarcasm can be detected in the following example: 7

(36) a. Dà ya-kè su-nàa tsòoron à kaamàa su, sun gudù.as it-be 3p-IPV fear imp.SBJ arrest 3p 3p.CPL flee'As it is the case they are afraid of being arrested, they fled.'

b. Dà su-kèe tsòoron à kaamàa su, ai sun gudù.as 3p-RI fear imp.SBJ arrest 3p well 3p.CPL flee'As they are afraid of being arrested, well [isn't it true that] they fled!'

c. Dà su-kèe tsòoron à kaamàa su, bàa sun gudù ba!as 3p-RI fear imp.SBJ arrest 3p NEG 3p.CPL flee NEG'As they are afraid of being arrested, isn't it the case that they fled!'

Sentence (36a), with a copular SSC, recalls a presupposed state of affairs (some people fear being arrested) and informs the listener about the consequent state of affairs. By contrast, in (36b) the speaker is sarcastic and the sentence can still be told even if hearer knows about both states of affairs and their connection. In this case, the function of (36b) would be to remind the hearer of a situation that is embarrassing (if it is his/her people) or funny (if hearer is unrelated to the people). In fact, sentence (36b) is most natural with the particle ai 'well, isn't it true that, of course, indeed, etc.'. As seen in (36c), reduced SSCs can occur with emphatic negation. The emphatic negation would not be natural with copular SSCs. It is not really the case that copular SSCs cannot be used to express speaker attitude, however, reduced SSCs seem to have specialized in this function. These two points are illustrated next:

(37) a. Dà ya-kè ka-nàa tsòoro-n-tà, ai bà kà cêe ma-tà koomii ba.as it-be 2ms-IPV fear-of-3fs well NEG CPL say to-3fs nothing NEG'As it is the case you are afraid of her, well, you said nothing.'

22

b. Dà ka-kèe tsòoro-n-tà, ai bà kà cêe ma-tà koomii ba.as 2ms-RI fear-of-3fs well NEG CPL say to-3fs nothing NEG'As you are afraid of her, well, you said nothing.'

(38) a. ??Dà ya-kè ka-nàa tsòoro-n-tà, kaa cêe ma-tà wani àbù?as it-be 2ms-IPV fear-of-3fs NEG CPL say to-3fs something'As it is the case you are afraid of her, did you said anything?'

b. Dà ka-kèe tsòoro-n-tà, kaa cêe ma-tà wani àbù?as 2ms-RI fear-of-3fs NEG CPL say to-3fs something'As you are afraid of her, did you said anything?'

Given the general content of sentences (37), both copular SSC and reduced SSC can have the sarcastic use. Note that the sentences express what happened in reality (i.e., hearer is afraid of someone and hearer did not say a thing). However, as shown in (38), the speaker can also explicitly encode his/her sarcasm by using a stylistic reverse polarity (it is known hearer did not say a thing). In that case, sentence (38b), with a reduced SSC, is more natural, as indicated. Sometimes, the reduced SSC is more specialized, for example when expressing the pride or honor of the speaker. In this situation, the copular SSC is unnatural, as seen in the following:

(39) a. *Dà ya-kè kin ƙaarè aikì-n-kì, (ai) naa biyaa kì!as it-be 2fs.CPL finish work-of-2fs well 1s.CPL pay 2fs'As it is the case you finished the work, well, I paid you.'

b. Dà ki-kà ƙaarè aikì-n-kì, (ai) naa biyaa kì!as 2fs.RP finish work-of-2fs well 1s.CPL pay 2fs'When/as you finished your work, I paid you [if that need reminding]!'

In the background situation of sentences (39), both speaker and hearer are involved and agree on the factual events. In this sense, the basic information function of the copular SSC is irrelevant. However, situations may arise where the hearer (the worker paid) behaves in an inconsistent way (say, insinuating speaker is a bad payer or complaining not having enough available money, etc.). In that case, speaker may need to defend himself/herself and, for that purpose, the reduced SSC is the most natural option, while the copular SSC is simply unfelicitous, as indicated in (39a). Such sentences can also be used unprovoked, if a speaker just wants to be mean (for example in doing goorìi in Hausa, such as when one

23

helps and later reminds the receiver about it). The same pattern is observed in the case of scolding, as seen next:

(40) a. *Dà ya-kè su-nàa fìtaa, (ai) sai kà kiraa nì.as it-be 3p-IPV go.out well then 2ms.SBJ call 1s'As it is the case they are/were going out, you should have called me.'

b. Dà su-kèe fìtaa, (ai) sai kà kiraa nì.as 3p-RI go.out well then 2ms.SBJ call 1s'As they are/were going out, you should have called me.'

In (40), a potential causal event (people going out) could have as consequence the hearer's alerting the speaker. If hearer fails to act properly, then speaker can complain, in effect scolding or blaming the hearer. In this case, only reduced SSCs are adequate, as indicated in (40b). There may be more to say about the expressive functions of causal reduced SSCs, however, the purpose of this paper is not to be exhaustive in this regard. Furthermore, there may be a ranking between the various functions already surveyed here. This however will be left to future studies. 8

Besides backgrounding causal/reason clauses, reduced SSCs can also background consequent clauses and express what in Abdoulaye (2007a) was referred to as "profiling" (or "highlighting, prominence"). In profiling, the speaker emphasizes, contrasts, or requests overdue information (cf. Abdoulaye 2007a). In this regard, reduced SSCs are well distinct from copular SSCs, as shown in the following:

(41) a. Dà ya-kè Aishà ta-nàa fiddà kùjèeruu, bàaƙii zaa sù zoo (kèe nan).as it-be Aisha 3fs-IPV take.out chairs guests FUT 3p come this.means'As it is the case Aisha is getting chairs out, (this for me means) guests are coming.'

b. Dà Aishàta-kèe fiddà kùjèeruu, (àkwai mafaarii), bàaƙii zaa sù zoo (nèe).as Aisha3fs-RI take.out chairs there.is reason guests FUT 3p come cop.'As [if] Aisha is getting chairs out, (there must be a reason), [it is because] guests are coming.'

In (41a), the speaker infers a reason for Aisha's action (cf. discussion of (33b) above) and the inferential marker kèe nan 'this for me means' can be used. In (41b), the speaker assumes hearer cannot conceive of an explanation (a reason) for Aisha's action and supplies one (the event in the main clause). That the speaker indeed refers to the state of mind of the hearer is

24

shown by the presence in (41b) of the expression àkwai mafaarii 'there must be a reason', which can be used to prepare the hearer to accept an explanation. The expression àkwai mafaarii would be unfelicitous in sentence (41a). Similarly, instead of kèe nan 'this for me means', the copula nee/cee can naturally be used in (41b) to show that speaker implies hearer cannot arrive at the same conclusion on his/her own. There is a good indication that the profiling SSCs are distinct from all other SSCs and represent the most grammaticalized constructions. Indeed, it is only the profiling SSCs that can shed away the subordinator dà, as illustrated next:

(42) a. Bàaƙii zaa sù zoo (nèe), dà Aishà ta-kèe fiddà kùjèeruu.Guests FUT 3p come cop. as Aisha 3fs-RI take.out chairs.'[it is because] guests are coming, as [if] Aisha is getting chairs out.'

b. Bàaƙii zaa sù zoo (nèe) Aishà ta-kèe fiddà kùjèeruu.Guests FUT 3p come cop. Aisha 3fs-RI take.out chairs.'[it is because] guests are coming that Aisha is getting chairs out.'

(43) a. (Ai) Aishà taa fiddà kùjèeruu, dà bàaƙii su-kà zoo.well Aisha 3fs.CPL take.out chairs as guests 3p-RP come.'(Well), Aisha got the chairs out, when/as the guests came.'

b. *(Ai) Aishà taa fiddà kùjèeruu bàaƙii su-kà zoo.well Aisha 3fs.CPL take.out chairs guests 3p-RP come.'(Well), Aisha got the chairs out, when/as the guests came.'

In sentence (42a), the consequent reduced SSC (as in fact all other SSCs) can be postposed to the main clause. In this position, however, only the consequent reduced SSCs can drop the subordinator dà and be intonationally integrated with the main clause, as shown in (42b). Indeed in (43a), a causal reduced SSC can also follow the main clause but, as indicated in (43b), it is not possible to drop the subordinator dà and keep the SSC reading. The SSC in (42b) can be referred to as a "dà-less reduced SSC" and is the source of out-of-focus clauses, with which it has the same form (cf. Abdoulaye 2007a).

To summarize, the four logical possibilities in the combination "copular SSCs vs. reduced SSCs" and "causal/reason clause backgrounding vs. consequent clause backgrounding" have specialized functions. This section tried to give the general tendencies in this respect and it may not be forgotten that context is also important. For this reason, there may be an overlap between the constructions, as is normal in grammaticalization change. In fact, the functions expressed in the four cases can be better understood in the

25

framework of a grammaticalization scheme, starting from the causal/reason copular SSCs to the reduced SSCs, and branching in sub-schemes involving relative clauses and out-of-focus clauses. A provisional picture of the domain is given in Figure 1. 9

(Insert Figure 1 about here)

In Figure 1, the causal/reason copular SSCs occupy a central position. They stem out of an impersonal equational clause introduced by the temporal subordinator dà and develop in three paths to derive causal reduced SSCs, consequent copular SSCs and copular Constituent Specifying Clauses (CSCs). The CSCs are presupposed relative clause-like constructions that emphasize or contrast pronouns or nouns the referent of which is known by speech participants. When they apply to nouns with unknown referents, they primarily help in referent identification and so become relative clauses (cf. Abdoulaye 2004, 2007a). As shown in the Figure, both CSCs and relative clauses have copular and reduced form and are thus formally parallel to copular and reduced SSCs. In the Figure, consequent copular SSCs derive consequent reduced SSCs, which themselves derive the out-of-focus clauses found in constituent focus (cf. Abdoulaye 2007a). As indicated, a constant function in Figure 1 is presupposition marking, which characterizes all types of clauses (except the temporal relative clauses, cf. Abdoulaye 2007c). Besides this basic function, the constructions have specialized functions, as indicated. Only causal copular SSCs have a basic information function. Profiling (which encompasses emphasis, contrast, and information request) seems frequent and characterizes the most grammaticalized constructions. There are naturally alternative ways of conceiving the derivations (for example, causal reduced SSCs may derive consequent reduced SSCs and reduced CSCs). It is clear that further studies are necessary to establish the appropriate derivations and the mechanisms that would lead from one function to another (say, from information to expressivity or from speaker-based inference to clause profiling, etc., -or the alternative paths).

As is clear from the Figure, this paper posits a diachronic relation between the various constructions, even though they all coexist synchronically in the language. There isn't necessarily a conflict here. The diachronic aspect of Figure 1 lies in the claim that in the history of Hausa, there was a stage, Stage I, where only temporal clauses existed. Then there was a Stage II at which copular SSCs developed, and a Stage III at which reduced SSCs arose, etc. Indeed, it has been verified in numerous studies that earlier stages of a grammaticalization process persist with later stages (cf. the development of various uses of English let in Hopper and Traugott 1994: 12ff), so that typical grammaticalization changes

26

could be represented as "Stage I > Stage I/Stage II (> Stage II), as argued in Hopper and Traugott (1994: 120). In fact, in languages where enough diachronic data are available, one may find direct evidence for the proposed chronological development (cf. Hopper and Traugott 1994: 76, 92, 185f). Dialects of a language may also evidence different stages of the same grammaticalization process (cf. for example the various stages of the grammaticalization of verb zâa ‘start to go, be going’ to future in western and central dialects of Hausa, see Abdoulaye 2001; cf. also Heine and Claudi 1986: 82, 113, 126f for further examples).

To summarize, copular and reduced SSCs participate in a pattern of functions that show their connectedness. In particular, this section has established the fact that causal copular SSCs fulfill a basic information function while the other constructions all have specialized inferential, stylistic, or pragmatic functions.

5. ConclusionThis paper has shown that there are good indications that in Hausa, reduced SSCs, with the relative marking, derive from copular SSCs through clause merger. The paper uncovered formal and functional patterns that can be understood as grammaticalization processes from the periphrastic copular SSCs to the syntactically reduced SSCs. Indeed, the copular SSCs are more general and have reduced correspondents only in certain contexts. Functionally, copular SSCs are closer to the source temporal clauses and can accompany consequent main clause with a basic information function. By contrast, reduced SSCs specialize in conveying stylistic and pragmatic functions. In conclusion, one can say that in Hausa, the relative marking, as a tense/aspect alternation, arose in reduced SSCs after a clause merger process.

Notes* Hausa (Chadic) is spoken mainly in Niger and Nigeria. Primary data in this paper are from Katsinanci dialect and Standard Hausa. The transcription follows Hausa standard orthography with some changes. Long vowels are represented as double letters, low tone as grave accent, and falling tone as circumflex accent. High tone is unmarked. The symbol ɍ represents an alveolar trill distinct from a flap /r/. Written 'f' is pronounced [h] (or [hw] before [a]) in Katsinanci. A double question mark before an example indicates pragmatic oddness or mild ungrammaticality. The abbreviations are: 1, 2, 3 '1st, 2nd, 3rd person'; CPL 'Completive'; cop. 'copula'; df 'definite'; EVE 'Eventual'; f 'feminine'; FUT 'future'; HAB 'Habitual'; imp 'impersonal'; IPV 'Imperfective'; m 'masculine'; NEG 'negative'; p 'plural'; RI

27

'Relative Imperfective'; RP 'Relative Perfective'; s 'singular'; SBJ 'Subjunctive'; SSC 'scene-setting clause'.

This paper is part of a project on the relative marking in Hausa and other West African languages that was supported by the University of Antwerp Research Council through a postdoctoral research position at the Center for Grammar, Cognition and Typology, 2003-2004. I thank Johan van der Auwera for his detailed comments on various aspects of the project and for all the material support.1 That kèe is much older than nàa (and independent from it) is shown by the fact that kèe in contexts such as (5b), (6a), and (8a) normally does not take a prefixed subject pronoun, even if a noun subject is not specified (in that case the subject will be an independent pronoun). -Nàa on the other hand can take a prefixed subject pronoun in all contexts, although this prefixed pronoun is optional when a noun subject is specified.2 One-term identificational kèe nan has given rise to a host of secondary meanings and uses. Thus, kèe nan frequently marks restrictive/exhaustive meaning and the related 'it is all over/ you are set' use (cf. lìttàttàfân kèe nan 'these are all the books', cf. Abdoulaye 2006: 1162:n5, 2007a). Kèe nan also has the use 'that is just X', as in haalìn yâara kèe nan 'that is just how children are'. Kèe nan also means 'soon after' and marks temporal clauses carrying a fortuitous event that has no natural temporal relation with the event in the main clause, as seen in yaa tàfi Makkà kèe nan, sai akà yi juuyìn mulkìi 'soon after he left for Mecca, there was a putsch', where the traveler may have no connection with the putsch. Therefore, kèe nan is not quite equivalent to koo dà 'as soon as', contrary to the claim in Newman (2000: 559), as seen in koo dà ya tàfi Makkà, sai akà yi juuyìn mulkìi 'as soon as he left for Mecca, there was a putsch', where the travel is likely to have allowed the putsch. Finally, as seen in discussion of (33b), kèe nan has the inferential 'this for me means' use, which closely resembles the use of donc in French. In this function, it can appear at the beginning or end of a clause.3 As is clearly laid out in Figure 1, this paper proposes that temporal dà 'when'-clauses are the source of SSCs, which themselves are the source of relative clauses. It happens that in Hausa linguistic literature, simple temporal dà-clauses are though to derive from temporal relative clauses where the weak head would have been deleted (cf. Bagari 1976/87: 117, Jaggar 2001: 624, Newman 2000: 556, Tuller 1986: 113ff, Watters 2000: 223). This idea is based on data such as illustrated in the following:

28

(i) Yâaraa sun ga sarkii (lookàci-n) dà su-kà shìga gàrii.children 3p.CPL see emir time-df that/when 3p-RP enter town'The children saw the emir when they entered (i.e., visited) the town.''The children saw the emir when they were entering the town.'

In example (i), presence or absence of the word lookàcii 'time' seems to make no difference in the meaning of the sentence. In fact, the derivation of simple temporal clauses from relative clauses through head deletion is thought to be a general process in West African languages (cf. Wald 1987: 509n5, cf. also Reineke 1998: 103). However, there are many indications showing that this analysis may be inadequate and that simple temporal dà-clauses cannot be derived from temporal ‘time that’ relative clauses. First, there are indications, as suggested in the text, that temporal dà derives from comitative/instrumental dà. It can indeed take time nouns with an ambiguous interpretation (cf. ciiwòo yaa zoo dà dàamanaa 'the disease came with/ during the rainy season') and nominalized clauses (cf. dà jîn hakà... 'on hearing this...') that are equivalent to tensed clauses (cf. dà ta ji hakà... 'when she heard this...'). Many Hausa particles started as prepositions before becoming conjunctions. Secondly, simple temporal dà-clauses have an obligatory past time interpretation in all allowed TAMs, except for the Completive (but including even the Future I). The temporal relative clauses can have a past or future interpretation in all TAMs that they allow. It should also be noted that although the two clauses share some TAMs (general Imperfective, Relative Perfective, and Future I), only simple dà-clauses allow the Completive while only temporal relative clauses allow the Relative Imperfective and the Eventual. Another indication is that only simple temporal dà-clauses can express close consecutive events (cf. ta tàmbàyi Saanii. (*lookàcin) dà ya baa tà, sai ta tàfi 'she asked Sani [for something]. when he gave [it] to her, she went'). One may also note the existence of various time-related question strategies. One strategy uses the unanalyzable yàushè ‘when’ (cf. yàushè sukà zoo? ‘when did they come?’). Another strategy uses wànè lookàcii ‘which time’ (cf. wànè lookàcii sukà zoo? ‘which time [when] did they come?). Another strategy yet combines dà and wànè lookàcii (cf. dà wànè lookàcii sukà zoo? ‘at which time [when] did they come?’) or dà and yàushè (cf. dà yàushè sukà zoo? ‘at when did they come?’. Finally, the lookàcii-deletion analysis assumes that the semantically empty relative clause subordinator dà can acquire the more substantive meaning of 'when, as, because', which violates normal grammaticalization tendencies. In fact, if something should be deleted in a temporal lookàcin dà ‘time that’ clause, it would probably be the weaker grammatical element dà, since it is incorporated in relative pronouns (cf. wan-dà ‘one-that’) or is deleted

29

in profiling SSCs (cf. data (42b) below). For more details on some of these issues, see Abdoulaye (2007b).4 As indicated in the previous note, simple temporal dà-clauses admit the Relative Perfective, in which case they are surfacely indistinguishable from reduced SSCs. This is why sentence (1b) can have the reading 'when they got tired, they went under shade to rest'. In temporal dà-clauses, the Relative Perfective is a simple past marker (cf. its use in narrative storyline and main clauses of dialogical discourse) and does not contrast with the Relative Imperfective, so that sentence (1a) has no temporal interpretation and the general Imperfective must be used to convey the meaning 'when they were afraid of being arrested, they fled' (for more details on these issues, see Abdoulaye 2007c). Nonetheless, it should be noted that simple dà-clauses with the Relative Imperfective are apparently used to express a “circumstantial” meaning, which however is different from a simple temporal meaning, as illustrated next (sentence (ia) adapted from a radio broadcast):

(i) a. Dà ya-kèe màganàa dà ƴan jàɍiidàa ministà-n yaa cêe zaa àas 3ms-RI talk with journalists minister-df 3ms.CPL say FUT Isaakè duubà shirìn.repeat look.at program‘While speaking with journalists, the minister said the program will be reviewed.’

b. Dà ya-nàa/*ya-kèe màganàa dà ƴan jàɍiidàa, sai a-kà ɗaukèwhen 3ms-IPV/3ms-RI talk with journalists then imp-RP take.offwutaa haɍ sàu biyu.power till time two‘When he was speaking with journalists, power went off twice.’

Sentences such as (ia) are rare and belong to a rather formal register. The author has heard them exclusively in radio broadcasts (to translate the meaning ‘in/during a press conference…’). The adverbial clause with the Relative Imperfective here expresses the setting of the event in the main clause (i.e., ‘on the occasion/in the framework of a press conference…’). In contrast, for fortuitous events, as illustrated in (ib), regular Imperfective is required. There are indications that the adverbial clause in (ia) is in a frozen construction. For example, the adverbial clause cannot follow the main clause, contrary to the regular temporal clause in (ib). Also, the particle sai ‘then’ would be infelicitous in (ia). For these reasons, the adverbial clause in (ia) may be taken as a SSC, which has no copular

30

counterpart, just like other SSCs that are tightly integrated in their constructions (cf. note 7). The meaning contrast between (ia) and (ib) is probably related to the following contrast involving locative predications:

(ii) a. Dà ya-kè i-nàa can, su-nàa baa nì bàyaanìi kân aikì-n-sù.as it-be 1s-be.at there 3p-IPV give 1s report on work-of-3p‘As it is the case I was there, they kept giving me reports about their work.’

b. Dà ni-kè can, su-nàa baa nì bàyaanìi kân aikì-n-sù.as 1s-be.at there 3p-IPV give 1s report on work-of-3p‘As/while [because] I was there, they kept giving me reports about their work.’

(iii) Dà i-nàa can, su-nàa baa nì bàyaanìi kân aikì-n-sù.as 1s-be.at there 3p-IPV give 1s report on work-of-3p‘When I was there, they kept giving me reports about their work.’

The sentences in (ii), with a copular and a reduced locative SSCs, both assume that the reports were given only because the speaker was in a certain place and have hence a clear causal meaning. In contrast, sentence (iii) has no implication whether or not reports were given when the speaker was in other locations and has hence a simple temporal meaning.5 The complete affirmative paradigms for Completive, general Imperfective, Relative Perfective, and Relative Imperfective are given in Table (i) for reference.

Table (i): General and relative paradigms ("C" assimilates to next consonant)

Imperfective Relative Imperf. Completive Relative Perfective(all dialects) (others/ west) (all dialects) (others/ west)

1s inàa nikèe/ nikà naa na/ niC 2ms kanàa kakèe/ kakà kaa ka/ kaC2fs kinàa kikèe/ kikà kin kikà/ kinkà =kiC3ms yanàa = shinàa yakèe = shikèe/ shikà yaa ya/ yaC3fs tanàa takèe/ takà taa ta/ taC1p munàa mukèe/ mukà mun mukà/ munkà2p kunàa kukèe/ kukà kun kukà/ kunkà3p sunàa sukèe/ sukà sun sukà/ sunkàone anàa akèe/ akà an akà/ ankà

31

The relative Imperfective marker of central/east dialects derives from locative copula kèe, as suggested in Section 3.1. The locative copula may in turn have developed from a stance verb (on the development of verbal copulas in languages see Heine 1997: 72, 205 and Devitt 1990: 103). There are at least three proposals regarding the origin of the marker -kà (in the Relative Perfective, it is sometimes deleted or reduced to an assimilated consonant, depending on the dialect). Newman (2000: 571) thinks that it derives from a Proto-Chadic perfective marker *kà/kˋə, which in Hausa was restricted to narrative and other relative marking environments after the introduction of the new Completive. Schuh (2001: 5, 12) rejects such an analysis, objecting that as a perfective marker, -kà would not have been able to appear in western dialects in both perfective and imperfective environments. He proposes that -kà derives from a copula that used to mark focus. However, given the fact that presupposed SSCs - and not focus constructions- are the basic environment for relative marking, Abdoulaye (2007b) proposes that -kà was likely a former mood marker (probably marking a non-negative realis, cf. Abdoulaye 1997: 315f) and, as such, is compatible with both perfective and imperfective environments. Table (ii) shows how the current situation may have arisen.

Table (ii): Origin of the relative marker -kà

West dialects Central/east dialectsAspect + modal -kà Aspect + modal -kà

Perfective Imperfective Perfective ImperfectiveInitial *su[X]-kà *su[Y]-kà *su[X]-kà *su[Y]-kàCurrent sun-kà su-kà su-kà su-kèe

At the initial Proto-Hausa stage in Table (ii), after the suffixation of -kà, the distinction between the then-functioning perfective and imperfective would have been lost (distinction indicated by the "X" vs. "Y", whatever its exact locus was). This situation being untenable, western dialects adjusted the perfective form (by borrowing the Completive paradigm), while central/eastern dialects adjusted the imperfective form (by recruiting the auxiliarized copula kèe). The idea that -kà was some kind of addition to basic aspectual forms is also found in Jungraithmayr (1983: 227). However, he proposed that –kà was added only to the perfective, as a sequential marker.6 Another SSC-like construction uses the verb kasàncee 'be, turn out', as illustrated next:

32

(i) a. Dà ya kasàncee yâara sun ƙaarè aikì-n, an biyaa sù.as it turn.out children 3p.CPL finish work-df imp.CPL pay 3p'As it is the case the children have finished the work, they were paid.'

b. Dà yâara su-kà kasàncee sun ƙaarè aikì-n, an biyaa sù.as children 3p-RP turn.out 3p.CPL finish work-df imp.CPL pay 3p'As it is the case the children have finished the work, they were paid.'

In (ia), dà introduces an impersonal predication ya kasàncee 'it.RP turn out' that is followed by the inner adverbial clause. In the variant in (ib), the inner subject NP can replace the impersonal pronoun as subject of the higher predication. It is not possible to combine kasàncee 'be' with ya-kè 'it-be'. The kasàncee-clauses, like the copular SSCs, admit all TAMs except the pure irrealis (Subjunctive, Eventual, and Imperative). For other functions of kasàncee in relative clauses, see Newman (2000: 541).7 There are many syntactic contexts that allow only the reduced SSCs. For example, Hausa renders psyche-verb predications ('be surprised, be furious/annoyed, be glad, be pleased, etc.') with predications involving nominals, 'do surprise, feel annoyance, do happiness, feel pleasure, etc.', that are accompanied by SSCs expressing the triggering event. In these constructions, reduced SSCs seem to show more syntactic integration when they appear postposed in the psyche-predication. Indeed, in this position, reduced SSCs -but not copular SSCs- can integrate the predication as nominal complements. This is seen next:

(i) a. Naa yi màamaakìi dà su-kèe fìtaa dà wuri/ dà ya-kè su-nàa1s.CPL do surprise as 3p-RI go.out at early/ as it-be 3p-IPVfìtaa dà wuri.go.out at early'I am surprised that they go out early.'

b. Naa yi màamaakì-n dà su-kèe fìtaa dà wuri/ *dà ya-kè su-nàa1s.CPL do surprise-of as 3p-RI go.out at early/ as it-be 3p-IPVfìtaa dà wuri.go.out at early'I am surprised that they go out early.'

In (ia), where the SSCs are not syntactically bound with the noun màamaakìi 'surprise', both types are possible. However, when the word màamaakìi takes the possessive marker -n, indicating a noun complement structure, only reduced SSCs are possible, as shown in (ib).

33

Also, only reduced SSCs are able to appear as clausal subjects (cf. [dà sukèe fìtaa] bàa lâifii ba nèe '[that they go out] is not a problem'). 8 Reduced SSCs expressing causal/reason clauses can appear in isolation as a repartee to a previous speaker. In this context, they express derision, as illustrated next:

(i) a. Faɗà-n mìi nee ta-kèe yîi?noise-of what cop. 3fs-RI do'What is the matter her rumbling?

b. Wai dà yâara su-kèe wargii ƙoofà ɗaakì-n-tà.it.seems as children 3p-RI play door.of room-of-3fs'It appears it is just because children are playing before her room.'

c. Wai yâara su-nàa wargii ƙoofà ɗaakì-n-tà.it.seems children 3p-RI play door.of room-of-3fs'It appears it is because children are playing before her room.'

In reply (ib), the speaker, not only minimizes the reason with evidential wai 'it seems', but he/she also portrays the referent in a derisive way (i.e., referent should not be angry over a trivial, normal fact of life). By contrast, in (ib), the speaker only stresses the fact that he/she is not certain of the facts but there is no implication over the legitimacy of the reason. Another example is given next:

(ii) Ii, dà Abdù ya-kèe zuwàa hiiɍa wajen Fàati, kôo? (ai naa sanìi).yes as Abdu 3ms-RI going dating at Fati right well 1.CPL know'You want to say that Abdu is dating Fati, right? (well, I know that).'

Sentence (ii) can be uttered if speaker senses that an interlocutor is presenting the fact described as a secret or a hot news and goes by innuendoes. The speaker, getting the point, in effect signifies that the interlocutor's innuendoes are useless. It should be noted that copular SSCs cannot be used in (i-ii).9 Besides the basic constructions discussed in this paper, copular and reduced SSCs can combine with other particles such as haɍ 'till, to the point that, even', concessive koo 'even', tun 'since', etc. (cf. Abdoulaye 1997: 314). The complex interaction between these particles and the SSCs (as well as simple temporal dà-clauses) deserves further studies (cf. also Newman 2000: 505).

34

References

Abdoulaye, Mahamane L. (1992). Aspects of Hausa morphosyntax in Role and Reference Grammar. Unpublished Ph.D. dissertation, State University of New York at Buffalo.

-----. (1997). Presupposition and realis status in Hausa, Sprachtypologie und Universalien Forschung 50:308-328.

-----. (2001). The grammaticalization of Hausa zaa 'be going' to future. Journal of African Languages and Linguistics 22(1):1-32.

-----. (2004). Four uses of the relative marking in Hausa. Presentation at Syntax of the World's Languages, Max-Planck Institute for Evolutionary Anthropology, Leipzig, 5-8 August, 2004.

-----. (2006). Existential and possessive predications in Hausa. Linguistics 44(6):1121-1164.

-----. (2007a). Profiling and identification in Hausa. Journal of Pragmatics 29(2):232-269.

-----. (2007b). Deux types de propositions temporelles quand en Hausa. Bulletin de l'Institut des Sciences du Langage, University of Lausanne. (to appear)

-----. (2007c). Perfectivity and time reference in Hausa. ms. Abdou Moumouni University.

Bagari, Dauda Muhammad. (1976/87). Hausa subordinate adverbial clauses: Syntax and semantics, Imprimerie El Maarif Al Jedida, Rabat (PhD thesis, UCLA).

Bearth, Thomas. (1993). Satztyp and Situation in einigen Sprachen Westafrikas. In: Möhlig, Brauner, and Jungraithmayr, (Eds.), Afrikanistentag, pp.91-104. Cologne: Rüdiger Köppe Verlag.

Caron, Bernard. (1991). Le haoussa de l'Ader. Berlin: Dietrich Reimer Verlag.

Devitt, Dan. (1990). The diachronic development of semantics in copulas. BLS 16:103-115

Givón, Talmy. (1990). Syntax: A functional-typological introduction, Vol. 2. Amsterdam: John Benjamins.

Green, Melanie and Chris Reintges. (2003). Syntactic anchoring in Hausa and Coptic Wh-construction. In: Simpson, Andrew (Ed.), Proceedings of the Twenty-Seventh Annual Meeting of the Berkeley Linguistics Society, pp.61-72. Berkeley: University of California.

Güldemann, Tom. (2003). Grammaticalization. In: Nurse, Derek and Gérard Philippson, (Eds.), The Bantu languages, pp.182-194. London/New York: Routeledge.

35

Heine, Bernd. (1997). Possession: Cognitive sources, forces, and grammaticalization. Cambridge: Cambridge University Press.

----- and Mechthild Reh. (1983). Diachronic observations on completive focus marking in some African languages. In: Heine, Bernd et al., (Eds.), Sprache und Geschichte in Afrika (SUGIA) 5, pp.7-44. Hamburg: Helmut Buske.

----- and Ulrike Claudi. 1986. On the rise of grammatical categories. Berlin: Dietrich Reimer

Verlag.

Himmelmann, Nikolaus P. (1997). Deiktikon, Artikel, Nominalphrase: Zur Emergenz syntaktischer Struktur. Tübingen: Max Niemeyer.

Hopper, Paul and Elizabeth Closs Traugott. (1994). Grammaticalization. Cambridge: Cambridge University Press.

Hyman, Larry M. and John R. Watters. (1984). Auxiliary focus. Studies in African Linguistics 15(3):233-273.

Jaggar, Philip J. (2001) Hausa. Amsterdam: John Benjamins.

Jungraithmayr, Herrmann. (1983). Altes und neues Perfekt im Hausa und in anderen Tschadsprachen. Afrika und Übersee 66(1):221-230.

McConvell, Patrick. (1977). Relativization and the ordering of cross-reference rules in Hausa. Studies in African Linguistics 8(1):1-31.

Newman, Paul. (2000). The Hausa language: An encyclopedic reference grammar. New Haven: Yale University Press.

Newman, Roxana Ma. (1976). The two relative continuous markers in Hausa. Studies in African Linguistics, Supplement 6:177-190.

Reineke, Brigitte. (1998). Des constructions relatives dans les langues de l’Atakora. Cahiers Voltaïques 3:95-106.

Schachter, Paul. (1973). Focus and relativization. Language 49:19-46.

Schubert, Klaus (1971/72, 1972). Zur Bedeutung und Anwendung der Verbalparadigmen im Hausa und Kanuri. Afrika und Übersee 55:1-149, 208-227, 267-300; Vol. 56:90-118.

Schuh, Russell G. (2001). On the origin of the Hausa "relative" aspects and the "stabilizer", University of California Hausa webpage: http://www.humnet.ucla.edu/humnet/aflang/Hausa/hausa.html (page viewed on May 13th, 2005).

36

Thompson, Sandra and Robert Longacre. (1985). Adverbial clauses. In: Shopen, Timothy (Ed.), Language typology and syntactic description: Vol. 2 Complex constructions, pp.171-234. Cambridge: Cambridge University Press.

Tuller, Laurice A. (1986). Bijective relations in Universal Grammar and the syntax of Hausa. PhD thesis, University of California Los Angeles.

Wald, Benji. (1987). Cross-clause relations and temporal sequence in narrative and beyond. In: Tomlin, Russell S. (Ed.), Coherence and grounding in discourse: Outcome of a symposium, Eugen, Oregon, June 1984, pp.481-512. Amsterdam: John Benjamins.

Watters, John R. (2000). Syntax. In: Heine, Bernd and Derek Nurse, (Eds.), African languages: An introduction, pp.194-230. Cambridge: Cambridge University Press.

Wolff, Ekkehard H. (1993). Referenzgrammatik des Hausa. Müster-Hamburg: LIT.

37

Figure 1: Formal and functional clines in Hausa scene-setting clauses and derived constructions

Copular relative clauses Reduced relative clauses Temporal [Identification and emphasis] [Identification (and contrast)] relative clauses

Copular constituent specifying Reduced constituent specifyingclauses [Emphasis] clauses [Emphasis/contrast]

Temporal dà Causal copular SSCs Causal reduced SSCs+ presuppositional ya-kè [Basic information] [Speaker expressivity]

Consequent copular SSCs Consequent reduced SSCs Out-of-focus clauses[Speaker-based inference] [Clause profiling] [Constituent profiling]

[Presupposition] [Presupposition] [Presupposition]

1